This invention relates to a semiconductor-device manufacturing system, and more particularly to a system for processing a semiconductor substrate by plasma discharge used in, for example, a reactive ion etching (RIE) system with a high-frequency power supply.
In a dry etching system used in the process of manufacturing semiconductor devices, a substrate on whose surface a given mask pattern has been formed is placed in a vacuum reactive chamber. A reactive gas is introduced into the vacuum reactive chamber and at the same time, discharging plasma is generated, thereby causing reactive ions to etch the substrate.
At that time, high-vapor-pressure reaction products are generally produced as a result of the reaction between the reactive ions and the etched layer. The reaction products are exhausted. Depending on the pressure in the vacuum chamber, the type of reactive gas, the flow rate, and the amount of energy of the reactive ions, the rate of reaction with the etched film and the types of reaction products differ.
In a system for processing a semiconductor substrate by plasma discharge, one means for clearly verifying the presence or absence of the change of processing conditions with time and the degree of the change with time, if any, is to process a substrate in such a manner that it has a shape with a high aspect ratio.
The shape with a high aspect ratio is, for example, a contact hole, a via hole, or a trench. As a typical example, problems encountered in a case where a conventional dry etching system is used in the process of forming trenches for trench capacitors in the memory cells of, for example, a DRAM will be explained.
FIGS. 1A and 1B are sectional views of a substrate in the process of forming a trench for trench capacitor.
As shown in FIG. 1A, a TEOS (Tetraethyl orthosilicate) film 12 is first formed on an Si substrate 11 to be processed. Then, patterning is done to form a mask pattern, thereby forming a sample of the substrate.
Next, after each lot processing of semiconductor substrates by a magnetron RIE system, a sample of the substrate as shown in FIG. 1A is placed in a vacuum reactive chamber. Reactive gases HBr, O2, and NF3 are introduced into the vacuum reactive chamber at flow rates of 100, 10, and 70 sccm, respectively. Then, plasma discharge is effected at a pressure of about 200 mTorr (about 26.6 Pa) with a high-frequency power supply output of about 1000 W, thereby causing reactive ions to etch the sample.
As a result of this, a trench 13 for trench capacitor is formed at the Si substrate 11 as shown in FIG. 1B. Here, θ is the taper angle at the top of the trench 13 and D is the diameter of the bottom of the trench.
FIG. 2 shows the relationship between the number of lots of substrates processed by a conventional RIE system and the diameter D (μm) of the trench bottom. The number of substrates processed in one lot is, for example, 24 to 25.
As seen from FIG. 2, as the number of substrates processed increases, the diameter D of the trench bottom decreases. The reason is that, as the number of substrates processed increases, the degree of the taper at the top of the trench decreases, making the taper angle θ smaller gradually.
The cause of this is not clear, but the following phenomenon is considered to be taking place.
In processing a trench for trench capacitor, SiBrx, SiBryOz, and SiFα are mainly produced as reaction products. Although most of them are exhausted, part of them adhere to the relatively low-temperature parts of the vacuum chamber or decompose again into substances with lower vapor pressures and adhere to the inside of the vacuum chamber.
These deposits are estimated to be of the SiO2 family. When the deposits build up to form a film, they are exposed to degassing or plasma, which causes the film to decompose again. As a result, the actual flow rate of each process gas in the atmosphere in the vacuum chamber differs from the set flow rate, preventing the desired shape and etching rate from being achieved.
As described above, because the diameter of the trench bottom is closely related to the condition of the deposited film on the inside of the vacuum chamber, a grasp of the condition of the deposited film would help determine the time the inside of the vacuum chamber should be cleaned. It is, however, impossible to grasp the condition of the inner surface of the vacuum chamber from the outside.
At present, the standard value of the diameter D of the trench bottom for trench capacitor is 0.1 μm. In this situation, the vacuum chamber is opened to atmosphere and cleaned manually every, for example, eight lots on the basis of the data in FIG. 2. However, it is not clear whether the method is the best.
As described above, with the conventional dry etching system for manufacturing semiconductor devices, it is impossible to externally grasp the condition of the inner surface and others of the vacuum chamber. For example, in processing a trench for trench capacitor, the change of the diameter of the trench bottom with time dependent on the number of substrates processed is impossible to grasp and therefore the suitable cleaning time of the inside of the vacuum chamber cannot be determined.